Professional Documents
Culture Documents
127 Mcdonald2015
127 Mcdonald2015
Supplemental content at
IMPORTANCE Clostridium difficile infection (CDI) is associated with significant morbidity, jamainternalmedicine.com
mortality, and a high risk of recurrence. Proton pump inhibitor (PPI) use is associated with an CME Quiz at
initial episode of CDI, and PPIs are frequently overprescribed. For many, the use of PPIs could jamanetworkcme.com
likely be discontinued before CDI recurrence.
OBJECTIVES To determine whether PPI use was associated with a risk of initial CDI
recurrence, to assess what proportion of patients who developed CDI were taking a PPI for a
non–evidence-based indication, and to evaluate whether physicians discontinued
unnecessary PPIs in the context of CDI.
MAIN OUTCOMES AND MEASURES Recurrence of CDI within 15 to 90 days of the initial
episode.
RESULTS Using a multivariable Cox proportional hazards model, the cause-specific hazard
ratios for recurrence were 1.5 (95% CI, 1.1-2.0) for age older than 75 years, 1.5 (95% CI, 1.1-2.0)
for continuous PPI use, 1.003 (95% CI, 1.002-1.004) per day for length of stay, and 1.3 (95%
CI, 0.9-1.7) for antibiotic reexposure. The use of PPIs was common (60.7%), with only 47.1%
of patients having an evidence-based indication. Proton pump inhibitors were discontinued
in only 3 patients with CDI.
C
lostridium difficile is the most common cause of infec- for symptoms and were diagnosed by polymerase chain reac-
tious diarrhea in hospitalized patients1 and is associ- tion for toxin B. Patients who were polymerase chain reac-
ated with significant attributable costs (US $11 000 per tion positive without 3 episodes of diarrhea were excluded from
nosocomial case2) and morbidity and risk of mortality.3 Even the study. Exceptions included patients with ileus, toxic mega-
when an initial episode of C difficile infection (CDI) is success- colon, colitis determined by results from computed tomogra-
fully treated, approximately 20% of patients will experience 1 phy or colonoscopy, or autopsy-proven disease.
or more recurrences.4 The management of patients with mul- Patients within the cohort were evaluated for a recur-
tiple recurrences is challenging, and therapeutic options are rence of CDI, defined as recurrent symptoms (without alter-
limited. Therefore, when faced with a patient with an initial native explanation) in association with a second positive
episode of CDI, an important goal must be the prevention of C difficile polymerase chain reaction assay occurring between
subsequent episodes. 15 and 90 days after the initial episode. While the Society for
A reasonable approach to preventing recurrence is to mini- Healthcare Epidemiology of America and the Infectious Dis-
mize modifiable risk factors such as antibiotic exposure. Pro- eases Society of America guidelines suggest that a recurrence
ton pump inhibitor (PPI) use is prevalent and has been iden- should be defined as occurring within 56 days of the initial
tified as an important exposure before an initial episode of episode,19 they recognize that the definition is not evidence
CDI.3,5,6 Consequently, Health Canada7 and the US Food and based. We chose to include episodes occurring between days
Drug Administration8 have released advisories that in all pa- 57 and 90 because we believed that they were also clinically
tient populations PPIs should be used at the lowest dosage pos- important, as well as to ensure that our results would be
sible for the shortest duration that is clinically indicated. directly comparable to those of the conflicting large observa-
Despite associations with CDI and other recognized ad- tional studies15,16 that have previously addressed this issue.
verse events,8-11 PPIs are still overprescribed.12-14 Given that Censoring occurred at CDI recurrence or at death within 90
many patients receive these drugs without evidence-based in- days. Those who did not die or experience a recurrence were
dications, we hypothesized that unnecessary PPI use could be censored at 90 days.
discontinued in patients with CDI to minimize recurrence risk. There were 809 patients identified, of whom 55 were ex-
This would be an intervention of negligible cost with substan- cluded because they died before day 15 and therefore could
tial potential benefit. Nevertheless, although the association never have had a recurrence. This left a cohort of 754 patients
between PPI use and incident CDI is well established, the evi- whose cases were analyzed.
dence implicating PPIs with recurrence is limited to a few con-
flicting studies.15-18 Exposure Assessment
The objectives of our study were 3-fold. First was to de- The main exposure of interest was continuous PPI use. Expo-
termine whether PPI use was associated with a risk of initial sure was defined as the receipt of a PPI at the time of the ini-
CDI recurrence. Second was to assess what proportion of pa- tial episode of C difficile with either (1) ongoing exposure for
tients who developed CDI were taking a PPI for a non–evidence- at least 75% of days in the hospital in those who remained hos-
based indication. Third was to evaluate whether physicians dis- pitalized or (2) a discharge prescription for a PPI that was valid
continued unnecessary PPIs in the context of CDI. beyond 90 days after the initial episode. To ensure that pa-
tients with CDI recurrence remained exposed, they were also
reevaluated at the time of recurrence for a valid PPI prescrip-
tion from at least 48 hours before recurrence. More than 90%
Methods of exposed patients were receiving pantoprazole sodium (40
Ethics mg daily) because this is one of only 2 PPIs (along with dis-
Ethics approval was granted by the McGill University Health solvable lansoprazole) on our hospital formularies.
Centre Research Ethics Board. Informed consent was waived.
Other Demographics
Description of the Cohort We extracted information on the treatment of the initial epi-
This was a retrospective cohort study consisting of all pa- sode of CDI, the unit that the infection was attributed to (medi-
tients who developed an initial episode of health care– cal, surgical, critical care, or other), the admitting hospital (to
associated CDI between January 1, 2010, and January 30, 2013, control for differences in local colonization pressure), and
and who survived for a minimum of 15 days after their initial length of stay. In our hospitals, the duration of initial treat-
episode of nosocomial CDI. The study was conducted at 2 uni- ment of CDI is 14 days.
versity-affiliated hospitals, the 417-bed Montreal General Hos- Age, sex, and date of death (if deceased) were obtained
pital (Montreal, Quebec, Canada) and the 517-bed Royal Vic- from the patient registration database. Patient comorbidities
toria Hospital (Montreal, Quebec, Canada). were obtained from International Statistical Classification of
As per Quebec standards, health care–associated CDI cases Diseases, 10th Revision, data manually abstracted by medical
were those in which (1) symptoms occurred more than 3 days record personnel during the index admission (eAppendix in
after admission or (2) symptoms caused readmission in a pa- the Supplement). Vancomycin-resistant enterococci and
tient who had been hospitalized within the previous 30 days methicillin-resistant Staphylococcus aureus colonization sta-
and who was not a resident in a long-term care facility. All po- tus were obtained from the hospitals’ respective microbiol-
tential cases were adjudicated by an infection control nurse ogy laboratories.
jamainternalmedicine.com (Reprinted) JAMA Internal Medicine May 2015 Volume 175, Number 5 785
Table 1. Analysis of Risk Factors for Clostridium difficile Death or Recurrence Within 90 Days for 754 Patients in the Cohort
(continued)
786 JAMA Internal Medicine May 2015 Volume 175, Number 5 (Reprinted) jamainternalmedicine.com
Table 1. Analysis of Risk Factors for Clostridium difficile Death or Recurrence Within 90 Days for 754 Patients in the Cohort (continued)
Antibiotic Reexposure tant potential confounder included in all models. All variables
Ninety-nine percent (752 of 754) of patients were exposed to an- in the final model were checked for significant interactions.
tibiotics before the initial episode of C difficile. We obtained in- The proportional hazards assumption was tested by in-
hospital antibiotic reexposure data from our hospital phar- cluding a factor × time interaction for all variables. If this led
macy database following the standard 14 days of therapy for to violation of the proportional hazards assumption, the in-
index CDI (within 15-90 days of the initial episode). teraction with time was included in the model as the means
of correction of that violation, and the results for such vari-
Process of Evaluating Indications for PPIs Use ables are presented for days 15, 28, and 56.
A general internist (E.G.M.) reviewed the indications for PPI Data were missing for both initial antibiotic therapy and
use through a medical record review of all consecutive pa- antibiotic reexposure. For our multivariable analysis, we im-
tients admitted to a medical clinical teaching unit or to a criti- puted these variables with multiple imputation using chained
cal care unit (191 of 493 PPI users [38.7%]). One hundred of these equations.23 Imputation models included the variables to be
medical records were subsequently reviewed by a second phy- included in the analysis models. Twenty imputations were per-
sician (J.M.) for an independent verification of the indica- formed using logit for binary variables. Convergence was veri-
tion. κ Agreement was 0.94. Consensus for 3 cases of disagree- fied graphically, and the residuals were plotted against the fit-
ment was reached after discussion with a third physician ted values to evaluate model fit. A complete case analysis was
(T.C.L.). Patients admitted to these units (and not to surgical also performed to confirm that the conclusions were not
units) were chosen because they have detailed admission notes, changed by the imputation technique.
providing the most accurate history obtainable to determine
the rationale for PPI therapy (eAppendix in the Supplement).
Results
Statistical Analysis
With statistical software (STATA, version 13; StataCorp LP), we Among 754 eligible patients, there were 52 deaths without re-
used a Cox proportional hazards model to determine the cause- currence (6.9%), 193 documented recurrences (25.6%), and 509
specific hazards ratios for both initial recurrence and death as survivors without documented recurrence within 90 days
a competing risk. 20,21 We also performed competing risk (67.5%). The comorbidities and clinical characteristics of these
regression22 to determine subdistribution hazard ratios for both patients are listed in Table 1. Among 193 patients who experi-
end points, and we present both analyses for comparison.21 enced a recurrence, 157 recurrences (81.3%) occurred be-
We first conducted a univariate analysis looking for associa- tween days 15 and 56. The distribution of the timing of all 193
tions between comorbid conditions and C difficile recurrence. recurrences was as follows: 15 to 21 days (37 patients [19.2%]),
Factors that were associated with P < .25 and those believed a 22 to 28 days (43 patients [22.3%]), 29 to 35 days (28 patients
priori to be important were considered for inclusion in the mul- [14.5%]), 36 to 42 days (24 patients [12.4%]), 43 to 49 days (12
tivariable models. Backward elimination was then used to arrive patients [6.2%]), 50 to 56 days (13 patients [6.7%]), and 57 to
at the final models, with antibiotic reexposure being an impor- 90 days (36 patients [18.7%]).
jamainternalmedicine.com (Reprinted) JAMA Internal Medicine May 2015 Volume 175, Number 5 787
788 JAMA Internal Medicine May 2015 Volume 175, Number 5 (Reprinted) jamainternalmedicine.com
Table 3. Unadjusted and Adjusted CSHRs for Initial Relapse of Clostridium difficile and Estimated sHRs
for Recurrence Using Competing Risk Regression
jamainternalmedicine.com (Reprinted) JAMA Internal Medicine May 2015 Volume 175, Number 5 789
widely available and is subject to governmental restrictions demonstrating that ongoing nonindicated PPI use is safe, we sug-
in many countries. gest that the use of non–evidence-based PPIs should almost cer-
By contrast, PPI use is common, it is associated with CDI re- tainly be discontinued in the context of CDI.
currence in 2 large observational studies,15,16 and the associa-
tion is biologically plausible given the effect of PPI therapy on
the gut microbiome.37 For more than half of the patients with
incident CDI, we have shown that the indication for PPI use is
Conclusions
not evidence based and have demonstrated that their PPI could New drugs and fecal transplantation may represent future treat-
be safely discontinued. The cost of PPI discontinuation is neg- ments of CDI. In today’s era of choosing wisely, we propose that
ligible, and tapering the PPI dosage before discontinuation may the prevention of CDI recurrence should begin with the ces-
help to curb any complications such as symptoms of rebound sation of unnecessary PPIs, a potentially unnecessary and fre-
hyperacidity. In the absence of a randomized clinical trial clearly quently overused class of medications.
Published Online: March 2, 2015. 8. US Food and Drug Administration. FDA Drug 19. Cohen SH, Gerding DN, Johnson S, et al; Society
doi:10.1001/jamainternmed.2015.42. Safety Communication: Clostridium for Healthcare Epidemiology of America; Infectious
difficile–associated diarrhea can be associated with Diseases Society of America. Clinical practice
Author Contributions: Drs McDonald and Lee had stomach acid drugs known as proton pump guidelines for Clostridium difficile infection in
full access to all the data in the study and take inhibitors (PPIs). 2012. http://www.fda.gov/drugs adults: 2010 update by the Society for Healthcare
responsibility for the integrity of the data and the /drugsafety/ucm290510.htm. Accessed January 19, Epidemiology of America (SHEA) and the Infectious
accuracy of the data analysis. 2015. Diseases Society of America (IDSA). Infect Control
Study concept and design: McDonald, Frenette, Hosp Epidemiol. 2010;31(5):431-455.
Lee. 9. US Food and Drug Administration. FDA Drug
Acquisition, analysis, or interpretation of data: All Safety Communication: possible increased risk 20. Wolbers M, Koller MT, Stel VS, et al. Competing
authors. of fractures of the hip, wrist, and spine with the risks analyses: objectives and approaches. Eur
Drafting of the manuscript: All authors. use of proton pump inhibitors. 2012. http: Heart J. 2014;35(42):2936-2941.
Critical revision of the manuscript for important //www.fda.gov/Drugs/DrugSafety 21. Wolkewitz M, Cooper BS, Bonten MJ, Barnett
intellectual content: McDonald, Frenette, Lee. /PostmarketDrugSafetyInformationforPatientsandProviders AG, Schumacher M. Interpreting and comparing
Statistical analysis: McDonald, Lee. /ucm213206.htm. Accessed January 19, 2015. risks in the presence of competing events. BMJ.
Administrative, technical, or material support: 10. Health Canada. Proton pump inhibitors: 2014;349:g5060.
Milligan. hypomagnesemia accompanied by hypocalcemia and 22. Fine JP, Gray RJ. A proportional hazards model
Study supervision: Frenette, Lee. hypokalemia. 2011. http://www.hc-sc.gc.ca/dhp-mps for the subdistribution of a competing risk. J Am
Conflict of Interest Disclosures: None /medeff/bulletin/carn-bcei_v21n3-eng.php#_Proton Stat Assoc. 1999;94(446):496-509.
reported. _pump_inhibitors. Accessed January 19, 2015.
23. White IR, Royston P, Wood AM. Multiple
11. Health Canada. Plavix (blood thinner): new imputation using chained equations: issues and
REFERENCES recommendations for use with PPIs (antacids). 2011. guidance for practice. Stat Med. 2011;30(4):
1. Polage CR, Solnick JV, Cohen SH. Nosocomial http://www.healthycanadians.gc.ca/recall-alert 377-399.
diarrhea: evaluation and treatment of causes other -rappel-avis/hc-sc/2011/13545a-eng.php. Accessed
January 20, 2015. 24. Sung JJ, Chung SC, Ling TK, et al. Antibacterial
than Clostridium difficile. Clin Infect Dis. 2012;55(7): treatment of gastric ulcers associated with
982-989. 12. Maggio M, Corsonello A. Harmful effects of Helicobacter pylori. N Engl J Med. 1995;332(3):
2. Zimlichman E, Henderson D, Tamir O, et al. proton pump inhibitors: discrepancies between 139-142.
Health care–associated infections: a meta-analysis observational studies and randomized clinical
trials: reply. JAMA Intern Med. 2013;173(16): 25. Lau JY, Sung JJ, Lee KK, et al. Effect of
of costs and financial impact on the US health care intravenous omeprazole on recurrent bleeding after
system. JAMA Intern Med. 2013;173(22): 1559-1560.
endoscopic treatment of bleeding peptic ulcers.
2039-2046. 13. Zink DA, Pohlman M, Barnes M, Cannon ME. N Engl J Med. 2000;343(5):310-316.
3. Kwok CS, Arthur AK, Anibueze CI, Singh S, Long-term use of acid suppression started
inappropriately during hospitalization. Aliment 26. Chey WD, Wong BC; Practice Parameters
Cavallazzi R, Loke YK. Risk of Clostridium difficile Committee of the American College of
infection with acid suppressing drugs and Pharmacol Ther. 2005;21(10):1203-1209.
Gastroenterology. American College of
antibiotics: meta-analysis. Am J Gastroenterol. 14. Naunton M, Peterson GM, Bleasel MD. Overuse Gastroenterology guideline on the management of
2012;107(7):1011-1019. of proton pump inhibitors. J Clin Pharm Ther. Helicobacter pylori infection. Am J Gastroenterol.
4. Drekonja DM, Butler M, MacDonald R, et al. 2000;25(5):333-340. 2007;102(8):1808-1825.
Comparative effectiveness of Clostridium difficile 15. Freedberg DE, Salmasian H, Friedman C, 27. Lanza FL, Chan FK, Quigley EM; Practice
treatments: a systematic review. Ann Intern Med. Abrams JA. Proton pump inhibitors and risk for Parameters Committee of the American College of
2011;155(12):839-847. recurrent Clostridium difficile infection among Gastroenterology. Guidelines for prevention of
5. Janarthanan S, Ditah I, Adler DG, Ehrinpreis MN. inpatients. Am J Gastroenterol. 2013;108(11): NSAID-related ulcer complications. Am J
Clostridium difficile–associated diarrhea and proton 1794-1801. Gastroenterol. 2009;104(3):728-738.
pump inhibitor therapy: a meta-analysis. Am J 16. Linsky A, Gupta K, Lawler EV, Fonda JR, 28. Rostom A, Moayyedi P, Hunt R; Canadian
Gastroenterol. 2012;107(7):1001-1010. Hermos JA. Proton pump inhibitors and risk for Association of Gastroenterology Consensus Group.
6. Deshpande A, Pant C, Pasupuleti V, et al. recurrent Clostridium difficile infection. Arch Intern Canadian consensus guidelines on long-term
Association between proton pump inhibitor Med. 2010;170(9):772-778. nonsteroidal anti-inflammatory drug therapy and
therapy and Clostridium difficile infection in a 17. Kim JW, Lee KL, Jeong JB, et al. Proton pump the need for gastroprotection: benefits versus
meta-analysis. Clin Gastroenterol Hepatol. 2012;10 inhibitors as a risk factor for recurrence of risks. Aliment Pharmacol Ther. 2009;29(5):
(3):225-233. Clostridium-difficile–associated diarrhea. World J 481-496.
7. Health Canada. Proton pump inhibitors (antacids): Gastroenterol. 2010;16(28):3573-3577. 29. van Vliet EP, Steyerberg EW, Otten HJ, et al.
possible risk of Clostridium difficile–associated 18. Cadle RM, Mansouri MD, Logan N, Kudva DR, The effects of guideline implementation for proton
diarrhea. 2012. http://www.healthycanadians.gc.ca Musher DM. Association of proton-pump inhibitors
790 JAMA Internal Medicine May 2015 Volume 175, Number 5 (Reprinted) jamainternalmedicine.com
pump inhibitor prescription on two pulmonary 32. Pathak R, Enuh HA, Patel A, Wickremesinghe P. versus vancomycin for Clostridium difficile infection.
medicine wards. Aliment Pharmacol Ther. 2009;29 Treatment of relapsing Clostridium difficile infection N Engl J Med. 2011;364(5):422-431.
(2):213-221. using fecal microbiota transplantation. Clin Exp 36. van Nood E, Vrieze A, Nieuwdorp M, et al.
30. Lai KC, Lam SK, Chu KM, et al. Lansoprazole for Gastroenterol. 2013;7:1-6. Duodenal infusion of donor feces for recurrent
the prevention of recurrences of ulcer 33. McFarland LV. Alternative treatments for Clostridium difficile. N Engl J Med. 2013;368(5):
complications from long-term low-dose aspirin use. Clostridium difficile disease: what really works? 407-415.
N Engl J Med. 2002;346(26):2033-2038. J Med Microbiol. 2005;54(pt 2):101-111. 37. Amir I, Konikoff FM, Oppenheim M, Gophna U,
31. McFarland LV, Surawicz CM, Rubin M, Fekety R, 34. D’Agostino RB Sr, Collins SH, Pencina KM, Kean Half EE. Gastric microbiota is altered in
Elmer GW, Greenberg RN. Recurrent Clostridium Y, Gorbach S. Risk estimation for recurrent oesophagitis and Barrett’s oesophagus and further
difficile disease: epidemiology and clinical Clostridium difficile infection based on clinical modified by proton pump inhibitors. Environ
characteristics. Infect Control Hosp Epidemiol. 1999; factors. Clin Infect Dis. 2014;58(10):1386-1393. Microbiol. 2014;16(9):2905-2914.
20(1):43-50. 35. Louie TJ, Miller MA, Mullane KM, et al;
OPT-80-003 Clinical Study Group. Fidaxomicin
jamainternalmedicine.com (Reprinted) JAMA Internal Medicine May 2015 Volume 175, Number 5 791